Last Modified: Jul 12, 2023
Affected Product(s):
BIG-IP LTM
Known Affected Versions:
12.1.1, 12.1.0, 12.0.0
Fixed In:
13.0.0, 12.1.1 HF2
Opened: Jun 06, 2016 Severity: 3-Major Related Article:
K30592181
Unable to retrieve statistics or statistics are all 0 (zero) when they should not be zero. This is vCMP related. Guest virtual-disks always show in-use even when the guest is not in the running state. When the guest O/S is shut down, the GUI and TMSH do not show accurate information about status.
No statistics are available. Some statistics, such as traffic stats from TMM, will not be updated, though they may be non-zero. Others, such as system CPU stats that are calculated by merged, will be zero. This will be evident through all management interfaces such as TMSH, TMUI, SNMP, etc. vCMP guest O/S status is reportedly incorrectly.
If a directory is removed from /shared/tmstat/snapshots merged might run at 100% CPU utilization and become unresponsive.
If merged has stopped responding, restart the daemon using the following command: bigstart restart merged On a chassis with multiple blades or a device with vCMP guests, merged is running on each blade and on each guest. To determine which instance of merged is not responding, ssh to each blade and each vCMP guest, and run the following command to check the CPU utilization of merged: top -p `pidof merged` Any merged that has a CPU utilization of over 90% for more than a few seconds is potentially in this state and should be restarted. To prevent the issue from occurring, disable tmstat snapshots using the following command: tmsh modify sys db merged.snapshots value false.
The merged process no longer becomes unresponsive when a directory is removed from /shared/tmstat/snapshots.